I know a lot of people have heard of the connectomemapper or cmtk:
http://cmtk.org/connectomemapper/
I just wanted to let everyone know that i've taken the cmtk source code and made it a bit more BIAC friendly.
If you haven't, its a python pipeline for doing DTI connectome analysis with freesurfer, DTK, fsl tools. Previously they had minimal support for fMRI, but that's been changed ( at least at BIAC ). Basically your T1 goes through freesurfer, the DTI goes through DTK steps and freesurfer/fsl tools are used to register everything together. An 83 region ROI set is made, then there are various stats about the dti data are put into a connectome using these ROIs.
In the BIAC version i've replaced the dicom converter with a BXH converter module. Now you can start from BXH headers and are not forced to start with all dicom series, or generate your own nifti series and trick cmtk with fake dicoms. If converting a DTI it will extract the gradient directions for you and replace the path listed under "custom file". If converting an fMRI it will extract the TR for future use.
Also, i've integrated much of the resting_pipeline as an rsFMRI module. Previously they only ran mcflirt and gave the ROI timecourse. Now all theses steps from the resting_pipeline are included:
slicetime correction, motion-correction, motion regression, WM/Ventricle nuisance regression, bandpass filtering and DVARS scrubbing
I've added a couple calculations to the final connectomes that are produced:
-radial/axial diffusivity
-DTI fiber voxel count
-ROI timecourses
-rval/zrval stats from the fMRI connectome
So now, there's DTI and fMRI stats/vals/etc based on the same set of ROIs in subject space. I also took care to only include unique voxels into the various calculations.
The gui can be accessed with the command:
BIAC-connectomemapper
If you want to access the command line version, in python you'd first have to add the path before you imported the libraries:
import sys
import os.path
sys.path.insert(0,'/mnt/nodespace/common/python/biac-cmp/lib/python/')
Thus far i've only dealt with the freesurfer parcellation, so please ignore the "Lausanne2008" scheme for now. Most of the documentation from the cmtk site is still relevant, in the upgraded rsFMRI module its pretty straight forward, especially if you've used out resting_pipeline before.
The main result is a cff file ( connectome file format: http://cmtk.org/cfflib ) with all the connectomes/etc inside. I wrote a viewer named "cffviewer.py" that can display the various connectomes and timeseries (if showing an fMRI stat). All the outputs from freesurfer, dtk, etc are also there.
